Open JamesBonddu opened 2 weeks ago
it looks like that the milvus cluster is not healthy, but i don't know what did yo do before that?please share more info and the full milvus pod logs for investigation. Could you please refer this doc to export the whole Milvus logs for investigation?
/assign @JamesBonddu /unassign
看起来 milvus 集群不健康,但我不知道您之前做了什么?请分享更多信息和完整的 milvus pod 日志以供调查。您能否参考此文档导出整个 Milvus 日志以进行调查?
/分配@JamesBonddu /取消分配
我使用docker-compose部署的,如何导出这些日志. 目前milvus-cluster大概有4w多的collection, 每次重启 应该直接docker-compose up -d还是 先docker-compose down 然后再docker-compose up -d ,每次重启cluster会变成恢复状态 恢复成可查询状态需要很久,而且需要删除milvus的pulsar中的队列中积压的数据才能恢复。
looks like there is a failure in describeIndex as there are 40K collections, @yiwangdr could you please help to take a look
/assign @yiwangdr /unassign
@JamesBonddu Datacoord is not ready.
@JamesBonddu Datacoord is not ready.
- How much data do you have per collection?
- Please check if datacoord constantly restarts. Do you have enough resources (CPU/memory usage) on datacoord and datanode?
ech collection has 6000 entity
cpu 512 memory 1T
why not put everything in one collection with partition key features ? It doesn't a good design to have one collection with only 6000 data
Is there an existing issue for this?
Environment
Current Behavior
Expected Behavior
No response
Steps To Reproduce
No response
Milvus Log
cron_upload_warn.log
Anything else?
why datacoord list_index will report this error , what should i config for this